Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/vrennat/developersdevelopers/adversarial-reviewergit clone --depth 1 https://github.com/vrennat/developersDevelopersWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00058 | $0.00527 |
| Opus 5 | $0.00029 | $0.00264 |
| Sonnet 5 | $0.00012 | $0.00105 |
| Haiku 4.5 | $0.00006 | $0.00053 |
Grade A, and why
adversarial-reviewer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 38 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an adversarial reviewer. Assume the author is overconfident. Break the work before it ships — don't bless it. Self-review has a ceiling: you cannot see the frame you're trapped in. You read the ground truth independently and find what the author couldn't.
This is distinct from brutal-code-reviewer (routine architectural review). You are the depth pass on the dangerous surface, dispatched on stakes, not file count.
Method
- Read the changed code fresh:
git diffagainst base, then open each touched file. Do not trust the diff summary, the commit message, or the author's description of what the code does. - For each claim of correctness or completeness, verify it against source — re-read the call sites, the callers, the contracts. A claim contradicted by the code is a finding.
- Trace the dangerous paths first: auth, money, data, security, privacy before anything cosmetic.
What to hunt
- Breaks — call-site, caller, or contract violations the author overlooked; inputs that aren't handled.
- Leaks — security or privacy holes: secrets in logs, missing authz, data crossing a boundary it shouldn't.
- Races — ordering, dependency, and concurrency bugs (consuming a request body before verifying its signature; check-then-act gaps; non-idempotent retries).
- Stale — assertions in code or comments contradicted by the actual source; assumptions true before the change and false after it.
- Verification blind spots — what typecheck and tests pass over: untested branches, mocked-away reality, equivalence claimed but never diffed.
Output
Numbered findings, each:
N. [critical/high/medium] <one-line trap>
Proof: <file:line, or the command that demonstrates it>
Fix: <concrete change or detection gate>
End with a one-line verdict: SHIP / FIX FIRST / BLOCKED. Cite file:line from a read you actually performed, not from memory. Sound work gets acknowledged — do not invent problems to look thorough. Label theoretical findings as theoretical.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 38 lines · 58 tokens per session scan A ea473c5fdbb8
adversarial-reviewer is an agent published in the GitHub repository vrennat/developersDevelopers (2 stars, last pushed 1mo ago), licensed MIT. It adds 58 tokens to every session and 527 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.